摘要
在云计算环境下,对处理海量多媒体数据的作业以及任务调度与资源分配算法进行建模,在此模型下提出一种云计算环境下面向海量多媒体数据的动态任务调度算法.算法以云系统中海量多媒体数据文件的分块多副本存储形式为基础来规划并行处理任务,以文件块和副本的映射关系为特征对云系统中数据节点执行聚类,以已执行完毕任务的历史反馈信息为基础来动态调度未执行任务.实验结果表明提出的算法对提高系统资源利用率和负载均衡有显著效果.
this paper build a model for scheduling task and resource allocation in the cloud computing environment, then present a dy- namic task scheduling algorithm which consider the features of massive multimedia data in the model. The layout of parallel processing tasks in the algorithm is based on the blocks of massive multimedia data files that is stored in the form of multi-copy, cluster the data nodes in the cloud system by the mapping between file blocks and their copies, and take the historical feedback information of the completed tasks as a basis for scheduling tasks that is not running. Experimental results show that the proposed algorithm has a significant effect on improving system resource utilization and load balance.
出处
《小型微型计算机系统》
CSCD
北大核心
2013年第4期783-788,共6页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目(61170283)资助
广东省自然科学基金项目(10351806001000000)资助
深圳市基础研究计划项目(JC201005250045A)资助
关键词
云计算
海量多媒体数据
动态任务调度
云存储
cloud computing
mass multimedia data
dynamic scheduling
cloud storage
